Billing for lab tests, especially genetic tests, is a mess. Bill Kerr has ideas about how to clean it up.

CMS announced this summer that it had taken steps to stop more than $1.6 billion in potentially improper Medicare lab payments since the beginning of the Trump administration.

Bill Kerr, M.D., MBA, co-founder and CEO of Avalon Healthcare Solutions, is an expert on lab billing practices. His company works for payers, adjudicating claims from labs and handling other billing matters. Kerr spoke to Managed Healthcare Executive today about how lab tests — and particularly genetic tests — are billed and the claims adjudicated and his ideas about how coding and billing practices could be improved.

This transcript has been edited for clarity and length.

Genetic tests are billed under one or two Current Procedural Terminology (CPT). 81479 is used a lot.

That’s a big one for sure. There are a couple others. That's definitely one of them.

So every segment of CPT coding has what are called the nine-nine codes or seven-nine codes, which are the codes you use when you think you've done a procedure that is not represented by CPT. That's not uncommon.

The volume of lab tests exceeds any other area of medicine in the use of those not-otherwise-classified codes. So right off the bat, you can see that CPT has fallen way behind the innovation of the lab industry on genetics, years behind — not even months behind — years behind, and that's just a problem for the industry as a whole. It creates some friction. That's one element of it.

Secondly, though, when you have that much work happening under nonclassified codes, you also tend to get a lot of waste that happens because they're going to bill for a lot of things.

Not-otherwise-classified codes?

There are two different not-otherwise-classified codes. The seven nine is one that says, “I did molecular pathology.” That means I took tissue and I did some DNA testing on it. There are some others that say I did this many genes and I did even more genes, but I won't tell you which genes they are because they don't have a code for that.

There's also then the tendency to put all those together, and that generates, at least, wasteful billing. Some people would call it fraud.

In fact, Medicare recently said some of it is fraud and took a bunch of money back. That's where you get to the almost a billion dollars that they've been saying they are going after. It's a problem on multiple fronts to have this much work happening with no code.

Where I see it a lot [at Avalon] is that we push paper around and figure out what's going on, and then we do try to stamp out some of what's called stack coding.*

As you may know, a significant number of Medicare fiscal intermediates have adopted a unique code specification method. They use something that was developed called the Molecular Diagnostic Exchange. In 26, it may be up to 29, states you must bill with this extra code. You bill the 81479, but then you say, “Here's a Z code,” and that Z code has had coverage criteria attached to it by Palmetto GBA [a subsidiary of BlueCross BlueShield of South Carolina], which is one of the fiscal intermediaries. It's then been licensed by the others to say, “And now the coverage criteria we can attach to it is this.”

Because that's the other problem. How do you build coverage criteria when there's no consistent billing methodology? And so, Medicare struggles with that. All the payers struggle with that. And I live in a world of constantly trying to solve that.

And the amount of administrative waste and medical waste as a percentage of lab expense is pretty significant, particularly when you take it as a percentage of genetic testing.

In 26 states these Z codes further label these genetic tests for billing purposes. Is that a plain English translation?

I'm going to give you more detail than you're probably going to want to publish. McKesson, under their healthcare IT division, first developed the Z Code coding system. Change Healthcare bought it. Palmetto GBA licensed it and built a coverage criteria wraparound that's got quality, like who can do the test and when is the test useful, and they call that whole thing Molecular Diagnostic Exchange. Palmetto developed that. They now own the Z codes — they bought that from Change, so they own the whole thing, and they license it.

Every genetic test by each specific provider has its own code. Using that coding system, you can know not only what test but also who's doing it, which is really important when you think about the fact that after COVID, everybody who bought a PCR machine, when COVID testing decreased, they needed to figure out what to do. They all became genetic test providers. And many of them are not good.

The [Molecular Diagnostic Exchange] system is actually really good, in all honesty. Now, not all the rules they put around the coverage criteria are great for commercial, but it works pretty well for Medicare. The specificity of the system is great. The CPT is forever behind.

There's real opportunity to do something like Molecular Diagnostic Exchange and its coding system. It should become the genetic test management system in the healthcare system. It would remove a lot of friction and give a lot of companies the tools to manage what they want to cover and provide more tools for all of us who are trying to manage fraud, waste, and abuse.

Are you saying that this Z code system that Palmetto owns that's pretty darn good needs to be incorporated into the CPT code system in some way, shape or form?

Yes, and I can give you the analogy that makes this make sense. Are you familiar with the National Drug Classification (NDC) for drugs? It's very analogous. For years, and they still exist, CPT had the J codes. But the industry always wanted to use NDC, which was available at the PBM platform level and the pharmacy level but not the medical benefit level because of CPT. They always wanted to use that, and so for years people started saying, “Populate the NDC into a text field on the claim,” and people would try to build rules around it. And eventually, they got to the point that they said the NDC code now gets its own data field on the claim, not a text field, and now everybody can use NDC coding to adjudicate drugs, regardless of whether they're covered on the medical claim or the pharmacy benefit.

You need the same thing here. The Z code is to genetic testing what NDC is to drugs. You need it to be adopted as an official classification system, in my opinion. You need it to be put into the data fields against which you can adjudicate and then become part of the post-payment record that can be mined for data analysis and auditing and things like that.

Chris Klomp** is actually pretty well versed in some of this stuff, surprisingly, and has been, I think, part of the reason Medicare has done a good job of going after fraud, waste, and abuse in the lab space, and in particular in the genetic space. He comes out of a healthcare tech background.

We've encouraged him to do this very thing for Medicare. But I would personally love to see it become standardized because we need something that is precise and specific to allow the industry to more effectively leverage genetic testing.

We've talked about the friction in the system. I'm going to give you one more data point on the problem. There are labs that love to do genetic testing that they can bill under 81479 or something similar for a few reasons. One I've told you is they can stack the codes. *

Another reason labs want to do stuff [under a code such as 81749] is they want to just collect as much genetic test information as possible because building up genetic test databases is useful because you can sell it to pharma. That's fine. But it's a way to get insurance companies to cover what they might not otherwise intend to cover because it may be newer or experimental. I think there’s a lot of waste built in administratively, and things not intended to be covered at all for certain members are now being paid for by the member and the payer. That gets solved if you have specific coding.

Isn't there an inherent problem with specific codes? A provider, a physician, might want to do this panel of genes but add one more, and it's a different test. It just seems like the variability in this area would be huge.

Let's say for noninvasive pregnancy testing, I'm going to do testing for hemoglobinopathies such as hemophilia, sickle cell, etc., and then the doctor says yes, but I also want to do, say, a test for cystic fibrosis. They just order the two different Z codes, so it becomes pretty easy because each one has a Z code. It could also be bundled to a third Z code, which represents all of it.

*Stack coding is using multiple billing codes for the same test.

**Chris Klomp is a top aide to HHS Secretary Robert F. Kennedy Jr. Klomp has been nominated to be deputy HHS secretary.